Ensembles

This is the module for ensemble meta-algorithms which use ELM as base learners.

NELM as base learner

Bagging ELM

Bagging implementation for Neural Extreme Learning Machine.

class ensemble.BaggingNELM

Bases: algorithm.NELM

Bagging ensemble of Neural regularized Extreme Learning Machine

D = None

None

alpha = None

None

fit(trainData, trainTarg, parameters)

TrainData:Data matrix n x m, with n instances and m features. TrainTarg:Target J-encoded matrix n x j. Parameters:Structure with the cross validated hyperparameters.

get_indicator(testPatterns)

TestPattern:Data matrix n x m, with n instances to predict and m features.

k = None

None

trainTargDecoded = None

None

AdaBoost ELM

AdaBoost implementation for Neural Extreme Learning Machine.

class ensemble.AdaBoostNELM

Bases: algorithm.NELM

AdaBoost ensemble of Neural regularized Extreme Learning Machine

D = None

None

alpha = None

None

fit(trainData, trainTarg, parameters)

TrainData:Data matrix n x m, with n instances and m features. TrainTarg:Target J-encoded matrix n x j. Parameters:Structure with the cross validated hyperparameters.

get_indicator(testPatterns)

TestPattern:Data matrix n x m, with n instances to predict and m features.

trainTargDecoded = None

None

Boosting Ridge ELM

Boosting Ridge AdaBoost implementation for Neural Extreme Learning Machine.

class ensemble.BRNELM

Bases: algorithm.NELM

Boosting Ridge ensemble of Neural regularized Extreme Learning Machine

alpha = None

None

fit(trainData, trainTarg, parameters)

TrainData:Data matrix n x m, with n instances and m features. TrainTarg:Target J-encoded matrix n x j. Parameters:Structure with the cross validated hyperparameters.

fit_step(y_mu, s)

Same part

predict(testPatterns)

TestPattern:Data matrix n x m, with n instances to predict and m features.

y_mu = None

None

Negative Correlation ELM

Negative Correlation AdaBoost implementation for Neural Extreme Learning Machine.

class ensemble.AdaBoostNCNELM

Bases: ensemble.AdaBoostNELM

AdaBoost Negative Correlation ensemble of Neural regularized Extreme Learning Machine

fit(trainData, trainTarg, parameters)

TrainData:Data matrix n x m, with n instances and m features. TrainTarg:Target J-encoded matrix n x j. Parameters:Structure with the cross validated hyperparameters.

fit_step(weight, pen, trainTarg, eye_matrix, s)

Weight matrix

get_indicator(testPatterns)

TestPattern:Data matrix n x m, with n instances to predict and m features.